Full-stack AI engineer, 6+ years shipping AI products end to end. Co-founded laccuna, an AI/data-science startup that delivered 12+ production ML systems for Fortune 500 clients — 2M+ predictions/month across manufacturing, media, finance, and telecom. Today I build agentic LLM applications for cybersecurity: on-prem tool-calling agents grounded in live security logs.
An AI-powered SOC analyst platform:
- 🛡️ Full-stack SOC dashboard integrating NetWitness, designed modularly for multi-SIEM compatibility
- 🗣️ Plain-English → SIEM queries via an on-prem tool-calling agent (Qwen on Ollama) — minutes of hand-written query syntax down to one sentence
- 🔒 Read-only, least-privilege agent execution with query validation and full prompt/response audit logging
- 📄 Automated PCI-compliance & incident reporting
| Project | Impact |
|---|---|
| AI SOC dashboard with agent chat — React · FastAPI · LangChain · Ollama/Qwen | Tier-1 analysts query SIEMs in plain English |
| BERT article-relevance classifier — PyTorch · HuggingFace | 82% F1 · analyst workload 100 → 15 |
| Distributed time-series forecasting — Ray · PyTorch · GCP | Forecasting pipeline for a US media company |
| Customer churn prediction — gradient boosting · GCP | −15% churn for a telecom |
| ML workflow engine on Kubernetes — FastAPI · Celery · GKE | Auto-scaling MLOps microservices |
Client work lives in private repos — happy to walk through any of it.
languages Python · TypeScript/JavaScript · SQL · Bash
ai / ml PyTorch · Transformers · LangChain · Ollama · RAG · LLM agents & tool calling · OWASP LLM Top 10
frontend React · Next.js · Tailwind CSS
backend FastAPI · Node.js · Celery · RabbitMQ · Nginx
cloud GCP (GKE · Cloud Run · MIG) · AWS · Docker · Kubernetes · CI/CD
data PostgreSQL · Pandas · Ray · Plotly
ai tooling Claude Code · Cursor · MCP